Method for determining whether a vehicle control command would interfere with future safe vehicle maneuvering
By determining the safety set and predicting the future vehicle status, the problem of avoidance of the vehicle when the SAM is not allowed to be executed in a safe manner is solved, and the execution of SAM in a safe state is realized, and the safety and transportation efficiency of the system are improved.
Patent Information
- Application Number
- CN201980096072.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2019-05-08
- Publication Date
- 2024-11-29
- Estimated Expiration
- 2039-05-08
AI Technical Summary
The prior art is difficult to determine how a vehicle avoids an undesirable vehicle situation in a state where situation avoidance manipulation (SAM) is not allowed to be performed in a safe manner.
By obtaining one or more safety sets, it means that the vehicle state range of the future SAM can be initialized with a successful prospect, and predict the future vehicle state based on the current vehicle state and control command, comparing the predicted future vehicle state with the safety set, and determining Control whether commands hinder future SAM.
It is achieved to ensure that the vehicle can perform situation avoidance manipulation in a safe state without affecting transportation efficiency, and improve the safety and robustness of the system.
Smart Images

Figure CN113795417B_ABST
Abstract
Description
Technical Field
[0001] The present disclosure relates to a method, a control unit, a vehicle and a system for determining whether a vehicle control command for controlling a vehicle associated with a current vehicle state prevents a future situation avoidance maneuver of the vehicle due to certain interference implementations.
[0002] The present invention can be applied in heavy vehicles such as trucks and construction equipment. Although the present invention will be described primarily with respect to semi-trailer type vehicles, the present invention is not limited to this particular vehicle, but can also be used in other types of vehicles, such as, in unit frame trucks, construction equipment and recreational vehicles. Background Art
[0003] Autonomous and semi-autonomous vehicles use various types of processing systems and sensor input signals for navigation and vehicle control. Advanced driver assistance systems (ADAS) are also based on sensor input signals. If a part of a vehicle system fails, it may be necessary to transform the vehicle into a minimum risk situation. This type of maneuver is called a minimum risk maneuver (MRM). MRM is an example of a more general situation avoidance maneuver (SAM), in which the vehicle performs a maneuver to avoid an undesirable situation, such as a detected risk scenario or simply an inconvenient situation related to the vehicle.
[0004] For example, if certain critical sensor signals necessary for safe maneuvering of the vehicle are lost, then a SAM in the form of a safety stop maneuver may be required. A safety stop maneuver includes, for example, maneuvering the vehicle to the side of the road and stopping the vehicle in a controlled manner. Because the vehicle may be at least partially "blind" during a safety stop maneuver, control must sometimes be performed based on alternative input signals. One such alternative input signal is dead reckoning, where wheel revolutions are used to estimate, for example, distance traveled, speed, and acceleration.
[0005] US2018 / 0224851 A1 relates to the problem of performing a safety stop maneuver in the event of a loss of GPS signals. In the event of a failure of the primary positioning system, a position estimate based on dead reckoning is used.
[0006] Todoran et al., “Just-in-Time Emergency Trajectories: A Formulation Towards Safety in Autonomous Navigation,” 2018 IEEE / RSJ International Conference on Intelligent Robots and Systems (IROS), 3422–3429, October 1, 2018, deals with robotic navigation. Trajectories are generated under a set of constraints that allow the vehicle to deviate from its planned trajectory to a safe location at any time.
[0007] US20170158205 relates to a method for determining when ADAS or AD functions can be activated and when these functions should be deactivated. The method comprises: determining a safe driving state; and verifying whether the safe driving state is obtainable from the current driving state.
[0008] A problem associated with performing a SAM such as a safe stop maneuver is that the vehicle may be in a state that does not allow the SAM to be performed in a safe manner, i.e., with a prospect of success. For example, the vehicle may be traveling too fast to turn according to the required SAM track curvature.
[0009] There is a need for methods of determining the permissible vehicle state space such that undesirable vehicle situations can be avoided, which methods take the above-mentioned facts into account. Summary of the invention
[0010] An object of the present disclosure is to provide a method for determining whether a vehicle control command for controlling a vehicle associated with a current vehicle state interferes with a future situation avoidance maneuver (SAM) of the vehicle. The method includes: obtaining one or more safety sets, wherein each safety set represents a range of vehicle states in which a future SAM can be initialized with a prospect of success. The method also includes: obtaining a current vehicle state and a control command. The method then predicts a future vehicle state based on the current vehicle state and the control command, and compares the predicted future vehicle state with one or more safety sets. If the predicted future vehicle state is not included in any of the one or more safety sets, then the method determines that the control command interferes with the future SAM.
[0011] For example, the method can be used to monitor control commands to check whether SAM can be executed at any time with a prospect of success. Thus, it is possible to allow "anyone" to drive the vehicle within the limits provided by the system, which is an advantage. This allows the requirements on the driving control system of the vehicle to be relaxed, which is an advantage.
[0012] Advantageously, according to some aspects, the method can even be proven safe for a certain class of assumed vehicle dynamic models.
[0013] The disclosed method enables an autonomous driving system (ADS) to not need to resort to SAM every time the driving system proposes a trajectory that it does not know is safe. With the disclosed method, there can be a "middle ground" in which the system can defer a decision to perform, for example, an MRM or safety stop maneuver and still be safe, thus providing the nominal driving system with an opportunity to regain control of the vehicle. This helps improve transportation efficiency, which is an advantage.
[0014] According to various aspects, the SAM is at least partially defined by a predetermined longitudinal acceleration profile and an associated lateral control law. Thus, advantageously, the lateral control law can be conditioned on a given acceleration profile. This enables efficient processing, since, for example, a vehicle dynamics model that depends on vehicle speed can be assumed to be known when determining the lateral control law, which is an advantage.
[0015] According to various aspects, the SAM is at least partially defined by a target vehicle state range that the SAM is to achieve. Thus, the SAM is considered successful as long as the maneuver transforms the vehicle to a state included in the target vehicle state range. This enables a degree of freedom in determining the SAM, which is an advantage. Furthermore, by achieving a target state range, the associated control law becomes less restricted, which is an advantage.
[0016] According to various aspects, SAM is impeded in the presence of at least one interference sequence that cannot be safely compensated by the corresponding control law to reach the target vehicle state range. Therefore, SAM being impeded does not necessarily mean that there is no control law that transforms the vehicle into, for example, the target vehicle state range, because it depends on the implementation of the interference sequence. According to some aspects, being impeded means that there is an interference sequence that does not have a corresponding control sequence that brings the vehicle into the target vehicle state range. Therefore, even if SAM is impeded, there may be a control sequence that transforms the vehicle into the target state range due to some interferences, but not due to all interferences.
[0017] According to aspects, one or more safe sets are determined based on respective one-step-long robust controllable sets that are recursively modified into N-step-long robust controllable sets.
[0018] Advantageously, therefore, an efficient method of generating a security set is provided.
[0019] According to aspects, the one or more safety sets are obtained at least in part from a processing device arranged to generate the safety sets offline.
[0020] The advantage is that a portion of the required calculations can be performed offline, as vehicle computing resources may be limited in processing power or constraints on processing time. Also, reducing the number of real-time or online calculations results in easier evaluation of performance.
[0021] According to various aspects, the one or more offline generated safety sets are nominal safety sets generated based on a predetermined acceleration profile and a maximum steering rut curvature. Thus, the vehicle can select the offline generated nominal safety set corresponding to the current driving condition, thereby saving computing resources because the vehicle does not need to generate the nominal safety set online.
[0022] According to various aspects, the method includes adjusting one or more safety sets based on a maximum rut curvature value associated with a current driving scenario to generate a corresponding adjusted safety set. In this way, an existing or known safety set (such as a nominal safety set generated offline) can be adjusted for the current driving scenario, thereby improving the accuracy of the entire system. In addition, existing knowledge of the driving route can be used to adapt calculations in the system so that transportation efficiency is further improved.
[0023] According to various aspects, the prediction includes predicting future vehicle states based on a vehicle dynamics model associated with the vehicle. The vehicle dynamics model may be accurately determined, for example, based on a known acceleration or velocity profile of the vehicle.
[0024] According to various aspects, the vehicle dynamics model is determined at least in part as a linearized single track model, wherein additive noise is used to account for unmodeled vehicle dynamics. Advantageously, unmodeled vehicle dynamics are accounted for, as this enables additional system safety and robustness.
[0025] According to various aspects, the prediction includes predicting future vehicle states based on a bounded disturbance model. By assuming a bounded disturbance model, calculations can be simplified, which is an advantage.
[0026] According to various aspects, the prediction includes predicting the uncertainty associated with the future vehicle state. In this way, additional system robustness is provided, which is an advantage.
[0027] According to various aspects, the prediction includes predicting a drivable area associated with a future vehicle state. The drivable area may not always be easily deducible from the road geometry. An advantage is that the disclosed method can be applied to more complex drivable areas.
[0028] According to various aspects, the method includes adjusting the control command by an intermediate driving system control command in case the control command interferes with a future SAM. The intermediate driving system control command may be determined to not interfere with the SAM, which means that the vehicle may continue to operate based on the intermediate driving system control command. In this way, transportation efficiency is improved because triggering the SAM is sometimes avoided.
[0029] According to various aspects, the method includes triggering SAM in the event that the control command interferes with future SAM.Thus, advantageously, the disclosed method may also be used to actuate SAM commands, thereby providing safe vehicle operation.
[0030] According to some aspects, the SAM is an MRM. Advantageously, the disclosed method can be used to trigger an MRM and ensure that an MRM can be initiated at any time with a prospect of success.
[0031] A control unit, a computer program, a computer readable medium, a computer program product, a system and a vehicle associated with the advantages discussed above are also disclosed herein.
[0032] Generally, unless otherwise clearly defined herein, all terms used in the claims should be interpreted according to their common meanings in the technical field. Unless otherwise clearly stated, all references to "one / the element, device, part, device, step, etc." should be publicly interpreted as referring to at least one instance of the element, device, part, device, step, etc. Unless clearly stated, the steps of any method disclosed herein do not have to be performed in the exact order disclosed. When reading the attached claims and the following description, other features and advantages of the present invention will become apparent. Those skilled in the art will recognize that the different features of the present invention can be combined to produce embodiments different from the embodiments described below without departing from the scope of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS
[0033] With reference to the accompanying drawings, the following is a more detailed description of an embodiment of the present invention cited as an example. In the drawings:
[0034] Figure 1 A vehicle is schematically illustrated;
[0035] Figure 2 illustrates an example situation avoidance maneuver performed by a vehicle;
[0036] Figure 3 to Figure 4 schematically illustrates an example security set;
[0037] Figure 5 is a graph illustrating a velocity profile;
[0038] Figure 6 Diagram showing situation avoidance maneuvering lateral control;
[0039] Figure 7 to Figure 8 schematically illustrates an example vehicle state prediction operation;
[0040] Fig. 9 An example set of situation avoidance maneuvers is illustrated;
[0041] Fig.10 The adjusted security set is schematically illustrated;
[0042] Fig.11 is a flow chart illustrating the method;
[0043] Fig.12 a schematic diagram of a control unit;
[0044] Fig.13 An example computer program product is shown; and
[0045] Fig.14 A vehicle control system is schematically illustrated. DETAILED DESCRIPTION
[0046] The present invention will now be described more fully hereinafter with reference to the accompanying drawings, in which certain aspects of the invention are shown. However, the present invention may be embodied in many different forms and should not be construed as limited to the embodiments and aspects set forth herein; rather, these embodiments are provided by way of example so that this disclosure will be thorough and complete and will fully convey the scope of the invention to those skilled in the art. Like reference numerals throughout the specification represent like elements.
[0047] It should be understood that the present invention is not limited to the embodiments described herein and illustrated in the drawings; in fact, those skilled in the art will recognize that numerous changes and modifications are possible within the scope of the appended claims.
[0048] Figure 1 A schematic articulated vehicle combination 1 is shown, which includes a towing vehicle 2 and two towed vehicles 3, 4. The towing vehicle can be a common truck suitable for commercial highway use or a tractor with a turntable, but it can also be an off-road truck, bus or recreational vehicle. In the example shown, the first towed vehicle or trailer 3 is a dolly with a drawbar connected to the trailer coupling of the truck. The dolly is provided with two axles 7. The second towed vehicle or trailer 4 is a semitrailer provided with a kingpin 8 connected to the turntable of the dolly. This example shows a long vehicle combination of a common type, but other types of vehicle combinations with other types of towing vehicles and other types and numbers of towed vehicles can also be used. Different vehicle combinations can include trucks with common trailers, trucks with center axle trailers, trucks with dollies and semitrailers, tractors with semitrailers, tractors with B-links and semitrailers, tractors with semitrailers and common trailers, or tractors with semitrailer dollies and semitrailers.
[0049] In the vehicle combination shown, the effective wheelbase Leq1 of the towing vehicle, i.e. the truck, is the length from the front axle 12 of the truck to the imaginary axle 13. The effective wheelbase Leq2 of the first towed vehicle, i.e. the car, is the length from the drawbar joint of the car to the imaginary axle 6. The effective wheelbase Leq3 of the second towed trailer extends from the kingpin 8 to the imaginary rear axle 9 of the trailer 4.
[0050] The tractor can be provided with various autonomous or semi-autonomous driving functions, including processing circuits and sensor systems. For example, the vehicle may include the following combined Fig.12 The control unit 1200 is discussed in more detail.
[0051] The vehicle may be connected 15 to a processing device 14 arranged to perform off-line calculations to assist the vehicle control system. The connection 15 is preferably wireless, but may also be a wired connection or a connection via some storage module such as a hard drive.
[0052] The vehicle dynamics of the vehicle 1 depends on the speed of the vehicle. For example, for a constant turning radius R, the lateral or centripetal acceleration a y According to a y =v x 2 / R depends on the vehicle's longitudinal velocity v x and the turning radius R. Also, trailers behave differently at different speeds. For a slow-moving articulated vehicle 1, as discussed in US 9,862,413 B2, the trailers 3, 4 will cut inwards for a constant radius turn, thereby expanding the swept area inwards, i.e., towards the centre of the circle segment representing the turn. For a faster-moving articulated vehicle, the lateral acceleration a y The area swept by the articulated vehicle 1 is therefore generally increased at low speeds, i.e. a slowly moving articulated vehicle may actually sweep a larger area than the same articulated vehicle moving at a higher speed. For some "ideal" or "optimal" speed, the swept area is minimal, and then for sufficiently high speeds that slip occurs due to the increased centripetal acceleration, the swept area increases again.
[0053] Herein, a drivable area is an area in which a vehicle can be located without risk of damage to the vehicle or injury to the driver. The road itself is of course a drivable area as long as there are no obstacles on the road. However, depending on the situation, the shoulder of the road can be included in the drivable area, and depending on the situation and the type of vehicle, the area of the roadside can also be included in the drivable area. For example, if a relatively flat field extends to the roadside without a ditch in the middle, the flat field can be considered as a drivable area because an articulated vehicle can temporarily drive onto the field without significant risk of damaging the vehicle or injuring the vehicle occupants. Off-road vehicles such as dump trucks may be associated with a larger drivable area than ordinary cargo transport vehicles. The drivable area can be determined by the control unit 1200, which will be discussed in more detail below.
[0054] The vehicle state is a collection of variables that together describe what state the vehicle is currently in. Herein, the vehicle state includes variables associated with the vehicle's position (coordinates) and orientation (such as, for example, driving direction, steering angle, and articulation angle). The vehicle state also includes information associated with the vehicle's dynamic state, i.e., vehicle speed, acceleration, turn rate, etc. The vehicle state is typically represented as a vector of state variables x. As will be discussed in more detail below, the permissible vehicle state space can typically include both upper and lower limits on state variables, such as lateral position during maneuvering.
[0055] In this context, "safe" is to be given a broad interpretation. A safe vehicle state may be a state in which the vehicle and / or vehicle occupants and / or other road users are not subject to risk of injury or damage.
[0056] Some examples of safe and unsafe conditions can be given.
[0057] According to some aspects, a situation in which a risk of collision is not possible can be considered a safe state.
[0058] According to some other aspects, situations where a collision risk is not impossible, depending on the object, may still be considered a safe state. That is, a collision with a small bush or tree may be considered safe, while a collision with another vehicle or a larger object such as a brick wall may be considered unsafe.
[0059] According to some other aspects, a collision with another vehicle may be considered safe where, for example, via vehicle-to-vehicle (V2V) communications, it has been predetermined that the collision can be tolerated with low risk.
[0060] According to various aspects, a situation in which the vehicle risks leaving the drivable area is considered unsafe.
[0061] According to other aspects, as discussed above, depending on the characteristics of the ground outside the drivable area, it may be deemed safe to leave the drivable area.
[0062] The state space is an N-dimensional space representing the range of vehicle states. Physics always limits the state space depending on the type of vehicle, for example, by the maximum speed and acceleration that can be achieved. The control unit 1200 can impose other constraints on the state space, thereby limiting the vehicle state space to a certain range of values.
[0063] Autonomous and semi-autonomous articulated vehicles rely on sensor signal inputs to control the vehicle, whether or not there is a driver. The sensor systems arranged on the vehicle to support autonomous functions may include any of the following: radio detection and ranging (radar) sensors, light detection and ranging (lidar) sensors, vision-based sensors such as cameras, and global positioning system (GPS) receivers. These sensors monitor the vehicle's surroundings to detect obstacles and determine, for example, the geometry of the drivable area in front of the vehicle. The vehicle may also include a number of on-board sensors, such as steering angle sensors, articulation angle sensors, i.e. sensors that measure the angle between the towing vehicle and the trailer, wheel speed sensors, and inertial measurement units (IMUs).
[0064] If the vehicle loses sensor input from one or more sensor systems, a serious situation may result. For example, radar and lidar data sensors may crash, or the processing unit responsible for processing sensor signal data may suffer a power outage. If critical sensor signals are lost or some critical operation is otherwise impeded, then an automatic maneuver may be required, such as a safety stop maneuver. This maneuver can then be performed based on control using the on-board sensor systems. That is, dead reckoning using wheel speed and possibly steering wheel angle sensors.
[0065] Herein, a minimum risk maneuver (MRM) is a maneuver that transitions the vehicle into a safe state. A safe stop maneuver is an example of an MRM. However, an MRM may also include maintaining a constant speed in a given lane or performing an obstacle avoidance maneuver.
[0066] An MRM is an example of a Situation Avoidance Maneuver (SAM). The category of situation avoidance maneuvers includes all maneuvers that can be performed to avoid an undesirable situation, such as a detected risky situation. A SAM may also be related to, for example, a difficult parking situation.
[0067] In situations where, for example, the vehicle speed is too high or the vehicle is in an inappropriate lateral position relative to the road boundary, SAM may not always be possible to perform in a safe manner. If SAM is possible for all interference effects that may occur during the maneuver, then SAM is considered to be initialized with a prospect of success. On the other hand, if there is at least one interference event that cannot be compensated by the corresponding control command, then SAM success is not guaranteed. However, this does not mean that MRM cannot be performed for all possible interferences but only for some interferences. In other words, at any given time, the existence of a single safe maneuver that can be performed is a sufficient condition for safety. However, if the maneuver cannot be initialized with a prospect of success at the next decision point, then it must be initialized immediately. Immediate execution is always possible because the ability to perform a safe maneuver at this decision point is the criterion for not performing a safe maneuver at the previous decision point.
[0068] According to some aspects, initialization with a prospect of success means that the operation can be completed successfully with a high probability. According to some aspects, this probability can be considered to be close to 100%, that is, close to guaranteeing the success of the operation. According to other aspects, the probability is less than 100%. In other words, a threshold value can be configured that determines the probability or likelihood required for the operation to be referred to as being initialized with a prospect of success.
[0069] Figure 2 An example SAM 23 is shown. Vehicle 1 is traveling on a road 21 having two lanes. It is very likely that the left lane may include vehicles going in the opposite direction and should therefore be avoided. Vehicle 1 determines that SAM is needed at position (A), for example due to some type of system failure. SAM track 23 is determined to transition vehicle 1 into a safe state at position (B) where it can come to a complete stop at the roadside 22. SAM 23 is determined by an acceleration profile that describes how hard the vehicle should brake at different points along the track and a lateral control law that describes how the vehicle should turn in order to safely transition into a safe state at position (B).
[0070] In order to ensure vehicle safety, a "safety net" can be included in the vehicle control functionality. The task of the safety net is to monitor the vehicle control commands to ensure that the vehicle does not end up in a state where, for example, a safe blind stop cannot be guaranteed. Conceptually, this monitoring problem can be viewed as a forward reachability problem. Forward reachability analysis is typically a computationally demanding operation that must be performed online to account for various initial conditions and different types of disturbances.
[0071] The monitoring problem in this safety net concept can be summarized by the following question: If we apply the current control action, can we ensure or predict that a successful blind stop can be performed with some reasonable probability at the next decision moment?
[0072] If we have a set of states that guarantee a successful blind stop, referred to below as the safe set, we can use candidate control inputs and a model of the system to predict whether these states belong to the safe set at the next sampling time. If we can confirm the robust set membership of the state vector, then the control input is approved by monitoring. If the control input is not approved, then SAM can be initialized, such as a blind stop.
[0073] According to various aspects, the set of safe states may be expressed solely in terms of the states of the model, ie, for a certain safe set S, the safe behavior of the vehicle may be expressed as x∈S.
[0074] Intuitively, if the operational design domain (ODD) is simple enough (e.g., if it is limited to highway driving), then the safe state can be expressed in terms of the lateral position of the vehicle.
[0075] Executing a safety maneuver explicitly prevents execution of the transport mission until safe nominal control can be restored. However, safety maneuvers are generally not unique; at any operating point, there are several ways to avoid an accident, and knowing more than one way makes the system more robust to changes in traffic conditions.
[0076] Some details will now be given on how a safe set may be determined.
[0077] A polyhedron is defined as the solution set of a finite number of linear inequalities. If an inequality can be removed from the description of the polyhedron without changing the solution set, then the inequality is redundant. Likewise, if an inequality is not redundant, then it is necessary (or non-redundant). If all inequalities describing the polyhedron are necessary, then these inequalities constitute a minimal representation of the polyhedron.
[0078] A polyhedral linear system is a discrete-time linear system of the form x(k+1)=A(k)x(k)+B(k)w(k), where x and w represent state variables and exogenous disturbances, respectively. According to some aspects, the exogenous disturbances may be assumed to be bounded. According to some other aspects, the exogenous disturbances may be assumed to be within a certain range with a given probability, e.g., 5 9s or 99.999%. Polyhedral linear systems are generally known and will not be discussed in further detail herein.
[0079] refer to Figure 3, for a given target set X, a one-step robust controllable set (or pre-image set) S = Pre(X, W, Δ) is defined as the set of states that are robustly mapped to X by x(k+1) = A(k)x(k) + B(k)w(k), where Δ denotes the convex hull of the matrix pair (A, B) and w∈W. Pre-image sets are discussed, for example, by F. Borrelli, A. Bemporad, and M. Morari in "Predictive Control for linear and hybrid systems", Cambridge University Press, 2015, and therefore will not be discussed in more detail herein.
[0080] For all w∈W, the original image set S is a control signal u (shown in Figure 3 The interference set can be assumed to be bounded, or it can be assumed to represent a certain set with a fixed probability of occurrence.
[0081] If X can be defined using a set of linear inequalities, that is, if for some matrix H and some vector h, Then the one-step robust controllable set Pre(X, W, Δ) can be taken as Assessment, where The element j is given by Given, where we introduce the notation H for the j-th row of the matrix H j .: and introduces h for the jth element of vector h j Therefore, if the set W is a polyhedron, then the one-step robust controllable set can be computed by solving many linear programs (LPs).
[0082] Figure 4 The vehicle 1 is illustrated executing the SAM 23, which starts at time T0 in vehicle state x(T0). It is desired to transition the vehicle state x to a state included in the state set X in the period T1 to T0.
[0083] For a given target set X, which starts at time T0 and in which the desired vehicle state should be at a certain time T1, the N-step robust controllable set Pre of the system x(k+1)=A(k)x(k)+B(k)w(k) can be N (X, W, Δ) is recursively defined as
[0084] Ω0=X、Ω i+1 =Pre(Ω i ,W,Δ)∩X,Pre N (X, W, Δ) = Ω N .
[0085] Figure 5 A graph 500 is shown which illustrates the vehicle 1 in Figure 4 A predetermined acceleration profile 510 is shown to be followed during a SAM 23. According to various aspects, it is assumed that this velocity profile is known for any given SAM. Therefore, to determine the SAM, it is only necessary to determine the appropriate lateral control law based on the assumed predetermined acceleration profile 510.
[0086] For a certain t s , the N-step robust controllable set is the system x(k+1)=A(k)x(k)+B(k)w(k) with N=(T1-T0) / t s The N-step robust controllable set is also discussed in "Predictive Control for linear and hybrid systems" by F. Borrelli, A. Bemporad and M. Morari, Cambridge University Press, 2015, and therefore will not be discussed in more detail in this article.
[0087] An example algorithm for computing an N-step robust controllable set will now be provided. It should be noted that the algorithm consists of computationally demanding operations, namely, the evaluation of the pre-image set and the computation of the minimum representation. If the interference set is only hinted by a simple boundary, then the cost of evaluating the pre-image set can be significantly reduced.
[0088] The algorithm for computing the N-step robust controllable set can be given as follows:
[0089] Ω0=X,
[0090] For k = 1, ..., N, perform
[0091] -Ω k =Pre(Ω k-1 , W, Δ) ∩ X, and
[0092] - Find Ω k The minimum representation of .
[0093] The N-step robust controllable set is then represented by Pre N (X, W, Δ) = Ω N Given.
[0094] An example will now be given of how the safety set may be determined in practice by the vehicle 1. It should be appreciated that the present disclosure is not limited to this exact manner of determining the safety set. Rather, one skilled in the art will recognize that the safety set may be determined in many different ways using mathematical analysis, computer experiments, or even ad hoc investigations and actual experiments.
[0095] For this example, it is assumed that a linear model describing the lateral dynamics of the vehicle 1 with respect to the road or with respect to some estimated drivable area is available for the predetermined acceleration profile 510. It is further assumed that the vehicle state at time k is given by the state variable x(k). Then;
[0096] x(k+1)=A(k)x(k)+B(k)u(k)+D(k)w(k),
[0097] y(k)=C(k)x(k)+E(k)w(k),
[0098] z(k)=F(k)w(k),
[0099] where A models the vehicle dynamics, B models the effect of the control command u on the vehicle state x, D models the effect of the disturbance w on the vehicle state x, y represents the measurement of the vehicle state x via C and E models the effect of the disturbance w on the measurement y via E. The variable z represents the disturbance measurement via F, which is optional.
[0100] Figure 6 The SAM track 23 is illustrated from an initial state x(T0) at position (A) to a target state at position (B). Here, the example vehicle state x includes a three-dimensional position vector x, a velocity vector v, and an acceleration vector a. At time Ti, the actual vehicle state is x(Ti) represented by the triangle, which differs from the SAM track at time Ti represented by the star by an amount e(Ti). The goal of the SAM control system is to minimize the error e or some function of the error e. According to some aspects, the target values of both the velocity v and the acceleration a are predetermined by the SAM. Therefore, only the lateral control law is required to fully define the SAM track 23. Therefore, refer to Figure 6 , the longitudinal velocity profile during SAM can be predefined, and linear state feedback can be used to control the lateral motion during blind stops, i.e., u(k) = L(k)(y r -y(k))+L ff (k)z(k), where L is the feedback gain matrix that may vary over time, and y r is the output reference vector and L ff is the feed-forward gain vector associated with the disturbance measurement z.
[0101] It should be noted that the closed-loop system during a blind stop can be expressed as x(k+1)=A'(k)x(k)+B'(k)w'(k), where A'(k)=A(k)-B(k)L(k)C(k), B'(k)=B(k)L ff(k)F(k)-B(k)L(k)E(k)+D(k)B(k)L(k) and w'(k)=[w(k) T y(k) T ] T . Thus, for this example, these algorithms can be applied to determine the N-step robust controllable set discussed above.
[0102] Figure 7 The general idea behind the disclosed method is illustrated. Vehicle 1 is at a certain time T0-t s is associated with the vehicle state x(k|k), where t s Related to certain update cycles of the autonomous driving system or ADAS system in vehicle 1.
[0103] The slightly simplified notation x(k|k) is broadly interpreted herein as the vehicle state x at time k given data up to time k. For ease of notation, the prediction of the vehicle state given data up to time k is then simply denoted x(k+1|k).
[0104] Assume that the control system in vehicle 1 outputs the control command u 1 Then based on the control command u 1 and the vehicle dynamics model to predict the state of vehicle 1 at time k+1. This prediction is Figure 7 is shown as x(k+1|x(k), u 1 ), that is, given the vehicle state x at time k and the control command u 1 The vehicle state at time k+1 is the information of the vehicle state at time k+1. The predicted vehicle state x(k+1|x(k), u 1 ) is included in the safety set S, which means that for all possible disturbances w∈W, the SAM for transforming the vehicle 1 into a safe state can be successfully executed. Therefore, the control command u 1 , and the vehicle control can continue. However, for the control command u 2 , predict x(k+1|x(k), u 2 ) is not included in the safe set S. This means that there is at least one realization of a disturbance w∈W for which there is no control sequence that can be used to transition the vehicle into a safe state. This means that the SAM at time k+1 will be 2 However, this does not mean that it is impossible to transform the vehicle into X, since there may be many realizations of w∈W in which the transformation can be performed.
[0105] Therefore, it should be understood that “impeded” in this context does not necessarily mean that it is impossible to transition the vehicle 1 to the target state set X (e.g., 2This simply means that there are potential realizations of disturbances w∈W for which no control law can be formulated to transform the vehicle into the set of target states X.
[0106] Figure 8 Some aspects of predicting a vehicle state x are schematically illustrated. At time k, the vehicle state x(k|k) is associated with a vehicle state uncertainty metric 810 that describes the accuracy of the vehicle state estimate. A perfect state estimate is exactly the same as the "true" vehicle state, while a more realistic vehicle state estimate will differ from the true state according to some error distribution. Common ways to represent vehicle state uncertainty are through a covariance matrix that represents the expected error variation or through an uncertainty ellipse. Uncertainty can also be represented by a polyhedron or some more general boundary construction.
[0107] At time k, a control command u 820 is issued that will affect the vehicle state, and at time k + 1, a vehicle state uncertainty 840 is issued. The uncertainty measure can be compared to the set of safe states 830 at time k + 1. In this way, a more robust system can be obtained because the vehicle state uncertainty can be taken into account when determining whether a control command will interfere with a future SAM.
[0108] Now consider a vehicle such as Fig. 9 An example with two safe stops is shown. In this example, the transitions of the vehicle into these two safe stops serve as examples of SAM. Each safe stop is defined by a set of vehicle states around a nominal safe stop position. Longitudinally, the safe stop position can be reached using a predefined acceleration profile. Laterally, the safe stop position is reached using a predefined control law. Therefore, SAM is implicitly defined by the longitudinal acceleration profile and the lateral control law.
[0109] Fig.10 The nominal safety set S is shown to have been modified to an adjusted safety set S'. For a certain vector γ>0, assuming that the interference set W can be expressed as W = {w = [α T β T ] T |-γ≤α≤γ, Qβ≤r}. When evaluating the preimage set according to the above discussion, we denote μ i,j ≥0 and π i,j ≥0 are used for the Lagrangian dual variables corresponding to the inequalities α-γ≤0 and -α-γ≤0, respectively. It should be noted that due to the complementary slackness, the Lagrangian dual variables cannot be simultaneously (element-wise) non-zero. We now introduce the notation λ i,j =max(μ i,j , π i,j), where max(a, b) is a vector containing the element-wise maximum values of its arguments, and it should be noted that Then the pre-image set can be directly expressed as where Δ γ represents the possible deviation from the value of γ used for the evaluation of the original image set. In a similar manner, the sensitivity can then be propagated through the recursive process discussed above to express the adjusted safe set as
[0110]
[0111] It is slightly interesting to note that the choice of γ does not affect the orientation of the linear inequalities, but only their distance from the origin.
[0112] Note that if γ is updated, redundant inequalities may become necessary. Therefore, the minimal representation should be calculated carefully, and some inequalities that are nominally redundant may have to be kept in the description of the set.
[0113] For the form When the driving situation motivates different choices of γ, it is in principle simple to adjust the size of the safety set. However, when the size of the set is reduced, it should be ensured that the current one-step prediction of the vehicle dynamics is still contained in the safety set, such as Fig.10 In the following, we provide an explanation of how the maximum Δ can be calculated online if α is a scalar. γ Assume that the one-step prediction of the system dynamics is represented by the convex hull of a set of points, i.e., x(k+1)∈Co(x1,...,x p ).
[0114] We compute the residual of the inequality that defines the safe set for each such point, ∈ j =Qx j -r,j=1,...,p. For ∈ j element i, we can then calculate the element that will lead to (∈ j ) i = 0 γ Then, the maximum allowable Δ γ is obtained as the minimum of these results.
[0115] Fig.11 A flow chart is shown which illustrates a method summarizing the aspects discussed above. A method for determining a vehicle state x k|k The associated vehicle control command u for controlling the vehicle 1 k A method for determining whether a future situation avoidance maneuver (SAM) 23 of the vehicle 1 is impeded. The method comprises:
[0116] Get S1 one or more security sets S1, S2...S K , where each safety set represents the range of vehicle states within which a future SAM 23 can be initialized with a prospect of success,
[0117] Get the current vehicle status x of S2 k|k and control command u k ,
[0118] Based on the current vehicle state x k|k and control command u k And predict S3 future vehicle state x k+1|k ,as well as
[0119] The predicted future vehicle state x k+1|k With one or more security sets S1, S2, ... S K Compare S4, and if the predicted future vehicle state x k+1|k Not included in one or more security sets S1, S2, ... S K In any of the above, the control command u is determined k Hindering the future SAM 23.
[0120] As explained above, the disclosed method can be used to monitor control commands to check whether SAM can be performed with a prospect of success at any given time. Therefore, it is possible to allow "anyone" to drive the vehicle within the boundaries provided by the system, which is an advantage. This allows the requirements on the vehicle's driving control system to be relaxed, which is an advantage.
[0121] The disclosed method also enables an autonomous driving system (ADS) to not need to resort to SAM every time the driving system proposes a trajectory that it does not know is safe. There can be a "middle ground" where the system can defer the decision to execute SAM and still be safe, thus providing the nominal driving system with an opportunity to regain control of the vehicle. This helps improve transportation efficiency, which is an advantage.
[0122] According to some aspects, the future SAM 23 is to be updated at a predetermined number N of discrete time periods t s Therefore, the SAM may be a SAM to be initialized in the next update operation or at some time in the future. According to an example, N=1.
[0123] Alternatively, the SAM 23 may be at least partially defined by a predetermined acceleration profile 510 and an associated lateral control law. By fixing the acceleration profile, the lateral control law may be determined based on known vehicle dynamics, which simplifies the determination of the lateral control law.
[0124] The SAM 23 may also be defined, at least in part, by a target vehicle state range X that the SAM is to achieve.
[0125] According to some aspects, SAM 23 is impeded in the presence of at least one interference sequence that cannot be compensated by the corresponding control law to reach the target vehicle state range X. Thus, SAM being impeded does not necessarily mean that there is no control law that transforms the vehicle into, for example, the target vehicle state range, because it depends on the implementation of the interference sequence w. According to some aspects, being impeded means that there is an interference sequence w that does not have a corresponding control sequence that brings the vehicle into the target vehicle state range. Thus, even if SAM is impeded, there may be a control sequence that transforms the vehicle into the target state range.
[0126] According to some aspects, one or more security sets S1, S2, ... S K is determined based on a corresponding one-step robust controllable set recursively modified to an N-step robust controllable set. One-step robust controllable sets are discussed above. These types of sets are also discussed in "Predictive Control for linear and hybrid systems" by F. Borrelli, A. Bemporad and M. Morari, Cambridge University Press, 2015.
[0127] According to some aspects, one or more security sets S1, S2, ... S K It is determined based on the linear difference inclusion model.
[0128] According to some aspects, one or more security sets S1, S2, ... S K is obtained at least in part from a processing device 14 arranged to generate a security set offline. This processing device 14 is shown in Figure 1 The advantage is that a portion of the required calculations can be performed offline, as vehicle computing resources may be limited in terms of processing power or constraints on processing time. Also, reducing the number of real-time or online calculations results in easier evaluation of performance.
[0129] According to some aspects, the one or more offline generated safety sets are nominal safety sets generated based on the predetermined acceleration profile 510 and the maximum steering rut curvature. Examples of safety set generation are given above. It should be appreciated that the disclosed method is not limited to the exact examples given above.
[0130] According to some aspects, the method includes adjusting S11 one or more safety sets S1, S2, ... S based on a maximum rut curvature value associated with a current driving scenario. K To generate corresponding adjusted security sets S'1, S'2...S'K . Combined with the above Fig.10 The adjustment of security collection is discussed.
[0131] According to some aspects, the prediction includes predicting S31 a future vehicle state x based on a vehicle dynamics model associated with the vehicle 1 k+1|k The vehicle dynamics model may depend on the vehicle speed, which is why it is advantageous to assume a known acceleration profile when determining the SAM. In other words, for a given SAM 23, the vehicle dynamics model is at least partially determined by a predetermined acceleration profile 510 of the vehicle 1.
[0132] According to some aspects, the vehicle dynamics model is determined at least in part as a linearized single rutting model, where additive noise is used to account for unmodeled vehicle dynamics.
[0133] According to some aspects, the prediction includes predicting S32 a future vehicle state x based on a bounded interference model. k+1|k As described above, according to some aspects, the exogenous interference w can be assumed to be within a certain range with a given probability, for example, 5 9s or 99.999%. According to some other aspects, the interference w can be bounded by a set of linear inequalities.
[0134] According to some aspects, the prediction includes predicting S33 and the future vehicle state x k+1|k The associated uncertainty.
[0135] Combination of the above Figure 7 and Figure 8 The vehicle state prediction is discussed.
[0136] According to some aspects, the prediction includes predicting S34 and the future vehicle state x k+1|k The associated drivable area.
[0137] To ensure safety, not only the vehicle state x can be predicted, but also the drivable area. Much like the model for the vehicle, the drivable area model is also dynamic and contains states. The vehicle and drivable area states can be combined to determine if the vehicle is within the drivable area. Multiple drivable area models can be used simultaneously to approximate the real drivable area, e.g., lanes, shoulders, general paths.
[0138] An example model of the drivable area consists of a constant width corridor along the road centerline curvature. The dynamic model contains bounds on the curvature and its rate of change. The vehicle model can then be set with respect to the road coordinates, and the distance between the road and the vehicle can be used to define a safe state. In addition to the road centerline curvature, a more complex drivable area model can also contain a safe vehicle orientation (including articulation angle) relative to the road (this will take advantage of the known geometry of the truck).
[0139] According to some aspects, the method comprises: k In case of obstruction of future SAM 23, the command u' is controlled by the intermediate driving system k To adjust the S5 control command u k The following will combine Fig.14 Discusses aspects related to the control of the intermediate driving system.
[0140] According to some aspects, the method comprises: k S6 SAM23 is triggered if future SAM 23 is blocked.
[0141] Fig.12 Components of a control unit 1200 according to an embodiment of the discussion herein are schematically illustrated in terms of a number of functional units. This control unit 1200 may be included in the articulated vehicle 1. The processing circuit 1210 is provided using any combination of one or more of the following: a suitable central processing unit CPU, a multiprocessor, a microcontroller, a digital signal processor DSP, etc. capable of executing software instructions stored in a computer program product, for example, in the form of a storage medium 1230. The processing circuit 1210 may further be provided as at least one application specific integrated circuit ASIC or a field programmable gate array FPGA.
[0142] In particular, the processing circuit 1210 is configured to cause the control unit 1200 to perform a set of operations or steps, such as combining Fig.10 For example, the storage medium 1230 may store the set of operations, and the processing circuit 1210 may be configured to retrieve the set of operations from the storage medium 1230 so that the control unit 1200 performs the set of operations. The set of operations may be provided as a set of executable instructions. Thus, the processing circuit 1210 is thus arranged to perform the method as disclosed herein.
[0143] The storage medium 1230 may also include a permanent storage device, which may be, for example, any single one or combination of magnetic memory, optical memory, solid-state memory, or even a remotely mounted memory.
[0144] The control unit 1200 may further include an interface 1220 for communicating with at least one external device, such as an external processing device arranged to generate the safety set offline. Thus, the interface 1220 may include one or more transmitters and receivers, including analog and digital components and a suitable number of ports for wired or wireless communication.
[0145] Processing circuit 1210 controls the overall operation of control unit 1200, for example, by sending data and control signals to interface 1220 and storage medium 1230, receiving data and reports from interface 1220, and retrieving data and instructions from storage medium 1230. Other components of the control node and related functions are omitted to avoid obscuring the concepts presented herein.
[0146] Fig.13 A computer readable medium 1310 is shown, which carries a computer program, the computer program including program code means 1320 for executing when the program product is run on a computer. Fig.10 The computer readable medium and the code means may together form a computer program product 1300 .
[0147] Fig.14 A system 1400 for controlling a vehicle 1 is shown, the system comprising:
[0148] The nominal driving system unit 1410, or unit A, is configured to generate a vehicle control command u k ,
[0149] Situation Avoidance Maneuver SAM 23 Generator Unit 1420 or Unit B, is configured to generate one or more SAMs, and
[0150] The safety set checking and selection unit 1440 or unit D is arranged to determine the vehicle control command u k Whether it hinders the future situation avoidance operation SAM 23 of vehicle 1. This unit combines the above Figures 1 to 11 The principle operation is discussed.
[0151] According to some aspects, the system 1400 further comprises: an intermediate driving system unit 1430 or unit C, which is arranged to generate a control command u' k , which controls the vehicle 1 so that under the control command u k In the case where the future SAM 23 is obstructed, the vehicle 1 may initialize the future SAM 23 .
[0152] The different units will now be discussed in detail;
[0153] 1410, Unit A: The Nominal Driving System (NDS) was developed for transport tasks and is responsible for "normal driving", i.e., lane keeping, lane changing, starting and stopping, etc. In the current setup, a wide range of vehicle dynamic capabilities are used, with high availability, and the NDS is not constrained by availability or correctness requirements while still having the opportunity to steer the vehicle, which is a novel feature. Its output is the control signal u k , such as, for example, [curvature, acceleration].
[0154] 1420, Unit B: The SAM generator unit, also known as the Safe Maneuver Generation (SMG) function, is responsible for computing a number of safe trajectories, i.e., situation avoidance maneuvers, each of which is safe (relative to the traffic conditions) to execute starting at the next decision point. Trajectories are defined by positions in state space over time, and all of these trajectories are included in the SAM output.
[0155] 1430, Cell C: The Intermediate Driving System (IDS) is an optional component that improves transportation efficiency by avoiding unnecessary transitions to situation avoidance maneuvers. Its sole responsibility is to propose a trajectory to control the vehicle with minimal dynamic changes to the current operating mode so that the vehicle can initiate a situation avoidance maneuver at the next decision point. For example, this may correspond to following a lane at a constant speed.
[0156] 1440, Unit D: The Safety Set Check and Selection (SSCS) is responsible for selecting the safety control signal and transmitting it to Unit E. In addition, if u k safety, then SSCS should be compared to u' k Priority control u k , and if u' k Security, then prefer u' over any SAM k If u k and u' k If none of them are safe, then the SAM that was judged to be safe at the previous decision point should be selected from all SAMs and the corresponding control signal should be sent to unit E.
[0157] According to the example, the following actions are performed online:
[0158] - The nominal driving system (and optionally also the intermediate driving system) sends control signals to the safety set check and selection.
[0159] - The SAM generator unit monitors the traffic situation and sends valid safety stops corresponding to the predefined longitudinal acceleration profile and the offline calculated safety set to the safety set check and selection.
[0160] -Safety set check and selection uses the vehicle dynamics model and the control signals obtained from the nominal driving system to perform a robust one-step forward prediction of the state vector. If any of the safety sets corresponding to the safety set check and selection that have been obtained from the SAM generator unit for a valid safe stop contains a prediction, then the control signal obtained from the nominal driving system is sent to motion control 1450. If no prediction is contained in any safety set, a similar prediction is performed using the control signal obtained from the intermediate driving system. If any of the safety sets contains a prediction, then the control signal obtained from unit C is sent to motion control. If no prediction is contained in any safety set, then SAM is initialized (verified as safe at the nearest decision point by units B and D). In addition, if B fails to recommend a valid safety stop, then SAM is initialized (verified as safe at the nearest decision point).
[0161] Some or all of the units A to E may be included in the control unit 1200. Some or all of the units A to E may be connected 15 to the processing unit 14 discussed above.
Claims
1. A method implemented in a safety set checking and selection unit in a vehicle (1), the method being used to determine a vehicle state (x k|k ) associated with a vehicle control command (u) for controlling the vehicle (1) k ) hinders future situation avoidance maneuvers of the vehicle (1), the method comprising: Obtain (S1) one or more security sets (S1, S2, ... S K ), where each safety set represents a range of vehicle states that are capable of initiating a future situation avoidance maneuver with a prospect of success; Obtain (S2) the current vehicle state (x k|k ) and obtain the control command (u) from a nominal driving system unit outside the safety set checking and selecting unit k ); Based on the current vehicle state (x k|k ) and the control command (u k ) to predict (S3) the future vehicle state (x k+1|k ); The predicted future vehicle state (x k+1|k ) and the one or more security sets (S1, S2...S K ) is compared (S4), and if the predicted future vehicle state (x k+1|k ) is not included in the one or more security sets (S1, S2...S K ), then determine the control command (u k ) prevent the future situation from being manipulated; as well as In the control command (u k ) hinders the avoidance maneuver of the future situation, the intermediate driving system control command (u' k ) to adjust (S5) the control command (u k ), wherein the intermediate driving system control command (u' k ) controlling the vehicle with minimal dynamic changes to the current operating mode to enable the vehicle to initiate the future situation avoidance maneuver at a next decision point.
2. The method according to claim 1, wherein: The future situation avoidance maneuver is to be performed within a predetermined number N of discrete time update cycles (t s ) to avoid manipulation of the initialization state.
3. The method according to claim 1 or 2, wherein: The situation avoidance maneuver is defined at least in part by a predetermined acceleration profile and a lateral control law associated with the vehicle (1).
4. The method according to claim 1 or 2, wherein: The situation avoidance maneuver is defined at least in part by a target vehicle state range (X) to be achieved by the situation avoidance maneuver.
5. The method according to claim 4, wherein: A situation avoidance maneuver is impeded in the presence of at least one interference sequence which cannot be compensated by a corresponding control law in order to reach the target vehicle state range (X).
6. The method according to claim 1 or 2, wherein: The one or more security sets (S1, S2...S K ) is determined based on a corresponding one-step robust controllable set that is recursively modified to an N-step robust controllable set.
7. The method according to claim 6, wherein: The one or more security sets (S1, S2...S K ) is determined based on the linear difference inclusion model.
8. The method according to claim 1 or 2, wherein: The one or more security sets (S1, S2...S K ) is obtained at least in part from a processing device (14) arranged to generate the security set offline.
9. The method according to claim 8, wherein: The one or more offline generated safety sets are nominal safety sets generated based on a predetermined acceleration profile and a maximum steering rut curvature.
10. The method according to claim 1 or 2, comprising: Based on the maximum rut curvature value associated with the current driving scenario, the one or more safety sets (S1, S2, ... S K ) to generate the corresponding adjusted security set (S'1, S'2...S' K ).
11. The method according to claim 1 or 2, wherein: The prediction includes predicting (S31) the future vehicle state (x) based on a vehicle dynamics model associated with the vehicle (1). k+1|k ).
12. The method according to claim 11, wherein: The vehicle dynamics model is determined at least in part by a predetermined acceleration profile of the vehicle (1) for a given situation avoidance maneuver.
13. The method according to claim 11, wherein: The vehicle dynamics model is determined at least in part as a linearized single rutting model with additive noise to account for unmodeled vehicle dynamics.
14. The method according to claim 12, wherein: The vehicle dynamics model is determined at least in part as a linearized single rutting model with additive noise to account for unmodeled vehicle dynamics.
15. The method according to claim 1 or 2, wherein: The prediction includes predicting (S32) the future vehicle state (x k+1|k ).
16. The method according to claim 1 or 2, wherein: The prediction includes prediction (S33) and the future vehicle state (x k+1|k ) associated with the uncertainty.
17. The method according to claim 1 or 2, wherein: The prediction includes prediction (S34) and the future vehicle state (x k+1|k ) associated drivable area.
18. A computer-readable medium (1310) carrying a computer program (1320), the computer program comprising program code means for executing the steps of the method according to any one of claims 1 to 17 when the computer program is run on a computer or on a processing circuit (1310) of a control unit (1200).
19. A method for determining a vehicle state (x k|k ) is associated with a vehicle control command (u k ) hinders future situation avoidance maneuvers of the vehicle (1), the control unit (1300) being configured to perform the steps of the method according to any one of claims 1 to 17.
20. A vehicle (1) comprising a control unit (1300) according to claim 19.
21. A system (1400) for controlling a vehicle (1), characterized in that: The system comprises: A nominal driving system unit (1410) configured to generate a vehicle control command (u k ), a situation avoidance maneuver generator unit (1420), the situation avoidance maneuver generator unit (1420) being configured to generate one or more situation avoidance maneuvers, The intermediate driving system unit (1430) is arranged to generate a control command (u' k ), the control command controls the vehicle (1) so that in the control command (u k ) hinders the future situation avoidance maneuver, the vehicle (1) is capable of initiating the future situation avoidance maneuver, and A safety set checking and selection unit (1440) is arranged to determine the vehicle control command (u k ) hinders future situation avoidance maneuvers of the vehicle (1), wherein the system (1400) is configured to perform the steps of the method according to any one of claims 1 to 17.
Citation Information
Patent Citations
Method, computer program, storage medium and electronic control unit for operating a vehicle
US20170158205A1
Method and apparatus for controlling autonomous driving vehicle using dead reckoning
US20180224851A1
Method for assisting the reversal of an articulated vehicle
US9862413B2
Drive assistance device for vehicle and drive assistance method for vehicle
JP2010260473A
Method for minimizing automatic braking intrusion based on collision confidence
US20150210280A1
